T-1000, a 10-Stage Axial Flow Turbine

[Toymaker]

This build thread will detail the design and construction of a 10-Stage, axial flow, Parson’s type turbine. The steam generator (aka, boiler) for this turbine will copy the SES Boiler-Burner design developed in the mid 1970’s for automotive use and has proven steam output of 10 LPM at 1000 F (538 C) at 1000 psi, and will be covered in a separate thread.

This turbine will use water as the working fluid, and will therefore need multiple rotor-stator stages to extract as much of the steam’s energy as possible. Modern power plants use steam turbines with 20 or more rotor-stator pairs which allow the steam to expand around 800 time. These turbines are very energy efficient, but they are also quite large and complex. My end use needs a mobile engine, so I've made a compromise between greatest efficiency and simplest, smallest design; 10 rotor-stator pairs should extract the majority of the steam's energy.

The high temperature & pressure I plan to use requires extensive use of stainless steel; the turbine case, nozzle, and stators are all made from 304 stainless. Heavy grid-lines in the drawing below are 1 inch.

I chose to use a split case design as it allows for easy assembly & diss-assembly and also provides for good visual checks of blade clearance, but of course the split case will also be a challenge to properly seal.

The finished case is made up from 4 separate pieces; top & bottom high-pressure section, which is welded to the top & bottom low-pressure section.

High pressure case halves were machined from a single block of 304 stainless, which was cut in half.

Each half was machined on my CNC mill and lathe.

 

[ddmckee54]

Is this another part of your ORC turbine project, or is this something new? Out of idle curiosity, what's your design max pressure?

 

[Toymaker]

I deemed the low pressure case for the last 6 rotor-stator pairs to be too large to reasonably machine from a solid block, so I paid a machine shop to roll the two halves from 3mm thick 304 stainless plate; judging from the appearance of the two halves, the shop made extensive use of a hammer to attain the final shape.

Clear plastic pieces were super glued to the cone to allow for mounting in a 4-jaw lathe chuck. Below photo shows inside of the low pressure case halves; the 5 grooves help to retain the stators.

Photo below shows top and bottom case halves after high and low pressure sections were welded together.

Outside of finished case halves, not screwed together.

Keeping high & low pressure sections perfectly aligned during welding was achieved by making & fitting 4 aluminum discs into the stator grooves and then bolting everything together before welding. No one will ever accuse me of making beautiful welds, but they are strong.

 

[Toymaker]

Is this another part of your ORC turbine project, or is this something new? Out of idle curiosity, what's your design max pressure?

The T-1000 project is an all new design and build using water as the working fluid. I still believe the use of ORC working fluids have many advantages over steam, but using them safely, with zero leaks, and low risk of exposure to hazardous gases is currently beyond what my home workshop is able to provide. So, at least for now, I will switch to steam.

Max pressure is 1000 psi at 1000 F (538 C); which is why I'm calling it T-1000. As I noted in the first post, the steam generator (aka, boiler) for this project will copy the SES Boiler-Burner design which was developed in the mid 1970’s for automotive use and which has proven steam output of 10 LPM at 1000 F (538 C) at 1000 psi.

I will start a separate thread in the Boilers Forum to discuss the SES Boiler.

 

[ddmckee54]

Yup, sealing that thing up at that pressure WILL be a challenge.

 

[Steamchick]

I really enjoy your ambition, and your machining looks excellent! I could offer advice, but you are in a league well beyond my experience, so I shall simply watch and learn from you, and others who are capable of offering advice. I think your idea to use a known design of steam boiler is practical, as you want a working solution in one step, instead of designing and developing everything from scratch!
WELL DONE!
K2

 

[Steamchick]

Incidentally, my Dad worked in a Power Station - on the steam turbines. (Full annual overhaul). They used Hylomar - a blue jointing paste - developed by Rolls Royce for their turbines, suitable for the steam pressures and temperatures involved. A very thin smear on the joint was adequate, as the bolted joint (like between your 2 halves) was a decent flat surface. The Hylomar compresses and squeezes out to a very thin smear, but seals the machining marks on the surface to prevent steam escape. The pressure it must withstand is below the 1000psi as the 1000psi starts at the first nozzles, and drops as the steam is accelerated (expands) through the initial nozzles, and thence through the turbine.
Hylomar was used by many British engine builders of all types from the development in the 1960s, as it replaced many previous chemistries of jointing compound for sealing flat machined joints.
But it says it is good for >250deg.C. Your steam starts at 538C... but what temperature after initial expansion through the first injection nozzles? - Would I be wrong to guess that 20 expansion stages (10 dynamic, 10 static?) reduce the temperature in 20 near-equal steps to the exhaust temperature? (Probably - I think!).
K2
https://hylomar.com/product/universal-blue/

 

[ddmckee54]

He's got the high pressure end welded to the low pressure end. Won't he need something that will work for the high pressure end? Since it will see that temp and pressure?

 

[Toymaker]

He's got the high pressure end welded to the low pressure end. Won't he need something that will work for the high pressure end? Since it will see that temp and pressure?

Yes, I will need a sealant with a temperature rating at or above 538 C (1000 F). Fortunately, Rocal makes the perfect sealant for my needs: Steamseal is made for sealing flanges and threads, and is rated for 600 C and 193 bar (2800 psi), and it's not too expensive, $38 for 400g.

 

[Toymaker]

Incidentally, my Dad worked in a Power Station - on the steam turbines. (Full annual overhaul). They used Hylomar - a blue jointing paste - developed by Rolls Royce for their turbines, suitable for the steam pressures and temperatures involved. A very thin smear on the joint was adequate, as the bolted joint (like between your 2 halves) was a decent flat surface. The Hylomar compresses and squeezes out to a very thin smear, but seals the machining marks on the surface to prevent steam escape. The pressure it must withstand is below the 1000psi as the 1000psi starts at the first nozzles, and drops as the steam is accelerated (expands) through the initial nozzles, and thence through the turbine.
Hylomar was used by many British engine builders of all types from the development in the 1960s, as it replaced many previous chemistries of jointing compound for sealing flat machined joints.
But it says it is good for >250deg.C. Your steam starts at 538C... but what temperature after initial expansion through the first injection nozzles? - Would I be wrong to guess that 20 expansion stages (10 dynamic, 10 static?) reduce the temperature in 20 near-equal steps to the exhaust temperature? (Probably - I think!).
K2
https://hylomar.com/product/universal-blue/

You're correct in assuming that both pressure and temperature will drop off quickly after the steam exits the nozzles, meaning the only area that will see max temp and pressure is the steam chest. But as ddmckee54 pointed out, the turbine case is essentially one piece, so the sealant needs to work at max temp and pressures.

I've never used this product, but Steamseal seems to be perfect for my needs as it's made for sealing flanges and threads, and is rated for 600 C and 193 bar (2800 psi).

 

[Toymaker]

Machining the 21 nozzles has proven to be an interesting challenge as the process requires milling very narrow, deep channels into 304 stainless; I knew this would be a problem area as I drew the design in CAD, but I wanted to give it a try. The narrowest section is 0.0218" wide and 0.165" deep, which requires using a 0.5mm end mill.

I start with a 2mm diameter end mill to remove much of the metal around each nozzle.

The second pass uses a 1mm end mill to cut deeper into the nozzle entrance and exit areas. Depth of each cut is limited to 1mm as deeper cuts resulted in breaking the endmill, or slowing the feed rate.

The 3rd and final pass is proving to be the most difficult as it requires using a 0.5mm end mill. Depth of each cut is limited to 0.1mm and turtles walk faster than the feed rate. This is the area I'm currently working on so I don't have any pics or video to show you at this time. Presently, I'm waiting on a different batch of endmills to arrive before I can proceed.

Below photo shows the nozzle before the final 0.0218" wide channel is cut.

 

[ddmckee54]

I was going to say that machining those nozzles in stainless could be some kinda fun. But I guess you already found that out didn't you. You buying those 0.5mm end mills by the gross?

 

[Toymaker]

I was going to say that machining those nozzles in stainless could be some kinda fun. But I guess you already found that out didn't you. You buying those 0.5mm end mills by the gross?

This nozzle is my first experience milling such small features into 304. The nozzle for the ORC turbine used 6061 aluminum and was MUCH easier to make, in part because the smallest channel was just over 1mm meaning I could use 1mm end mills.

I have learned to use HRC55 and stay away from HSS when machining stainless. I would like to try HRC75 Solid Cermet, but haven't been able to find 0.5mm in that material.

 

[Steamchick]

If the final slot is parallel, would a slitting saw be practical? - Just a guess, and probably stupid, but it seems like a more robust tool for cutting a small slot?
Perhaps the 0.5 mill could then smooth the slot into the adjacent curves?
K2

 

[Toymaker]

If the final slot is parallel, would a slitting saw be practical? - Just a guess, and probably stupid, but it seems like a more robust tool for cutting a small slot?
Perhaps the 0.5 mill could then smooth the slot into the adjacent curves?
K2

Using a slitting saw is a nice thought,...but,....

If you go back and look at the nozzle's CAD drawing and the last photo in post #11 you'll see there is not enough room to use a slitting saw with a diameter large enough to cut 0.165" deep. A total redesign would be needed to use a slitting saw.